JRRD Volume 47, Number 7, 2010

Pages 617–628

Journal of Rehabilitation Research & Development

Influence of gravity compensation on kinematics and muscle activation

patterns during reach and retrieval in subjects with cervical spinal cord
injury: An explorative study

Marieke G. M. Kloosterman, PT, MSc;1–2* Govert J. Snoek, MD, PhD;2–3 Mirjam Kouwenhoven, MD;2–3
Anand V. Nene, MD, PhD;2–3 Michiel J. A. Jannink, PhD2,4
1
Center for Human Movement Sciences, University Medical Center Groningen, University of Groningen, Groningen,
the Netherlands; 2Roessingh Research and Development, Enschede, the Netherlands; 3Roessingh Rehabilitation Cen-
ter, Enschede, the Netherlands; 4Laboratory of Biomechanical Engineering, University of Twente, Faculty of Engineer-
ing Technology, Enschede, the Netherlands

Abstract—Many interventions in upper-limb rehabilitation INTRODUCTION

after cervical spinal cord injury (CSCI) use arm support (grav-
ity compensation); however, its specific effects on kinematics
and muscle activation characteristics in subjects with a CSCI
are largely unknown. We conducted a cross-sectional explor-
ative study to study these effects. Nine subjects with a CSCI
performed two goal-directed arm movements (maximal reach,
reach and retrieval) with and without gravity compensation.
Angles at elbow and shoulder joints and muscle activation
were measured and compared. Seven subjects reduced elbow
extension (range 1.8°–4.5°) during the maximal reaching task
with gravity compensation. In the reach and retrieval task with
gravity compensation, all subjects decreased elbow extension
(range 0.1°–11.0°). Eight subjects executed movement closer
to the body. Regarding muscle activation, gravity compensa-
tion did not influence timing; however, the amplitude of activa-
tion decreased, especially in antigravity muscles, namely mean
change +/– standard deviation of descending part of trapezius
(18.2% +/– 37.5%), anterior part of deltoid (37.7% +/– 16.7%),
posterior part of deltoid (32.0% +/– 13.9%), and long head
biceps (49.6% +/– 20.0%). Clinical implications for the use of
gravity compensation in rehabilitation (during activities of
daily living or exercise therapy) should be further investigated
with a larger population.
Key words: electromyography, goal-directed movements,
gravity compensation, kinematics, rehabilitation, robot-assisted
therapy, robotics, spinal cord injury, tetraplegia, upper limb.
Damage to the spinal cord causes loss of motor and
sensory function of the body parts below the level of the
lesion. In patients with a cervical spinal cord injury
(CSCI), the arm and hand function is affected to varying
degrees according to the level and completeness of the
lesion [1]. Compared with other spinal cord injury-related
impairments, improvement in upper-limb function is one
of the highest priorities in patients with a CSCI [2]. Exer-
cise therapy integrated in an intensive rehabilitation pro-
gram to learn or relearn motor functions is considered
very important in optimizing the remaining upper-limb
function [1,3]. Even in the chronic stage, intensive exer-
cise therapy positively affects upper-limb motor control
and functional abilities in patients with a CSCI [4].
Abbreviations: 3-D = three dimensional, ADL = activity of
daily living, CSCI = cervical spinal cord injury, MRC = Medi-
cal Research Council, sEMG = surface electromyography.
*
Address all correspondence to Marieke G. M. Kloosterman,
PT, MSc; Roessingh Research and Development, Roessingh-
sbleekweg 33b, 7522 AH Enschede, the Netherlands; +31-
53-487-5777; fax: +31-53-434-0849. 
Email: m.kloosterman@rrd.nl
DOI:10.1682/JRRD.2010.02.0014

617
618

JRRD, Volume 47, Number 7, 2010

Literature indicates that motor learning or relearning is Apparatus

influenced by several key elements: active movements, A mechanical, passive device called Freebal [17]
intensity of practice (frequency, repetitions, and duration), (University of Twente, Enschede, the Netherlands; now avail-
use of feedback, task specificity, goal-orientated practice, able commercially as ArmeoBoom, Hocoma; Volketswil,
and variation [5–6]. Exercise therapy based on these motor Switzerland) was used to counteract the effect of gravity
learning or relearning principles asks great physical effort on the upper limb (Figure 1). The device has two
from patients with a CSCI who have impaired upper-limb slings—one is applied at the elbow and the other around
function. We presume that during goal-directed move-
ments, a large part of the preserved muscle force is
required to hold the arm against gravity; consequently, less
muscle force is available to perform the actual movements.
To facilitate goal-directed arm movements during activi-
ties of daily living (ADLs) [7] or exercise therapy, thera-
peutic devices are often used to support the weight of the
arm (e.g., with the Swedish Help Arm [Kinsman Enter-
prises, Inc; West Frankfort, Illinois]). In the last decade,
several innovative therapeutic devices, including robotics,
have been developed to support the affected upper limb
during exercise therapy [8–9]. In these robotic devices, dif-
ferent treatment modalities have been implemented, such
as passive, active-assisted, and active-resisted movements
[8]; consequently, gravity compensation is incorporated in
the design [8–9]. Until now, the effect of gravity compen-
sation on motor control and functional abilities has mainly
been investigated in nondisabled elderly [10] and stroke
patients [11–15]. Although many applications in rehabili-
tation after a spinal cord injury include gravity compensa-
tion during ADLs or exercise therapy, the specific effects
on kinematics and muscle activation characteristics
(amplitude and timing) in patients with a CSCI are largely
unknown. A cross-sectional explorative study that meas-
ured kinematics and surface electromyography (sEMG)
during goal-directed movements with and without gravity
compensation was conducted to study the effects of gravity
compensation in subjects with a CSCI.

METHODS

Subjects
Nine subjects with a CSCI (at least 1 year since
injury) were recruited from a local rehabilitation center.
Inclusion criteria for participation were motor injury
level C5–C7 (cervical) and age between 18 and 65 years.
Exclusion criteria were extreme shoulder pain, contrac-
tures of the upper limb, and/or spasticity preventing per-
formance of the required tasks. All subjects were assessed
according to the standard neurological classification [16].
Figure 1.
Freebal device for gravity compensation of upper limb. Source: Stienen
AH, Hekman EE, Van der Helm FC, Prange GB, Jannink MJ, Aalsma
AM, Van der Kooij H. Freebal: Dedicated gravity compensation for the
upper extremities. In: Proceedings of the 2007 IEEE 10th International
Conference on Rehabilitation Robotics; 2007 Jun 13–15; Noordwijk
aan Zee, the Netherlands. Piscataway (NJ): IEEE Press. p. 804–8.
DOI:10.1109/ICORR.2007.4428517
 619

KLOOSTERMAN et al. Gravity compensation in cervical SCI

the wrist. Each sling is connected to an independent

adjustable spring by way of an overhead cable and pul-
leys. During the goal-directed movements, this system
enables a constant amount of support throughout the
three-dimensional (3-D) working volume, irrespective of
the position and orientation of the arm [17].

Procedures
During the measurements, subjects sat in their own
wheelchairs (one subject was not wheelchair-dependent
and sat on a normal chair) in front of a height adjustable
table. In the starting position, subjects sat with their fore-
arm flat on the tabletop, elbow flexed at 90°, and hand on
the starting dot. Subjects performed two goal-directed
movements with and without the Freebal:
1. Maximal reaching task. This task consisted of three
maximum reaches in front of the subjects, without
gliding the hand and arm along the tabletop.
2. Reach and retrieval task. Subjects were instructed to
move at their own comfortable speed between a start-
ing dot and target dot on the table for 30 seconds. Both
dots were 10 cm in diameter, and the distance between
the dots was 35 cm (Figure 2(a)).

Measurement and Data Analysis

Kinematics
Kinematics were recorded with a 3-D optical move-
ment tracking system with six cameras (Vicon Nexus
1.3.109, Oxford Metrics Ltd; Oxford, United Kingdom).
Reflective markers were placed on 10 bony landmarks of
the arm and trunk: processus spinosus of the seventh cer-
vical and eighth thoracic vertebra, incisura jugularis, pro-
cessus xiphoideus, acromioclavicular joint, medial and
lateral epicondyle, radial and ulnar styloid, and distal
head third metacarpal (Figure 2(b)). Six cameras at
100 Hz recorded the 3-D marker trajectories. The acro-
mion marker was used for estimating the glenohumeral
rotation center. Scapular motion was disregarded because
scapular motion was not likely to participate in the ante-
flexion movement if the angle of elevation remains
below 60°.
The marker trajectories were visually inspected for
recording errors and missing marker data. If one trunk
marker was missing, we replaced it using the Vicon
BodyBuilder model (Metrics Ltd; Oxford, United King-
dom). This model estimated the position of the missing
marker by the position of the other three markers. We
Figure 2.
(a) Tabletop with start and target dots and (b) experimental setup. S =
starting dot, T = target dot.
replaced missing marker trajectories over a short period
(less than 10 samples) by linear interpolation. If data
were missing for longer periods or at the end of the reach
or retrieval movement, the movement cycle was removed.
Marker position data were converted to limb seg-
ments data according to the guidelines of the Interna-
tional Society of Biomechanics [18]; thereafter, joint
angles were calculated with Euler rotation. The elbow
joint angle (Figure 3(a)) was specified as the angle
between the longitudinal axis of the upper arm and the
forearm (full elbow extension was defined as 0°; forearm
perpendicular to upper arm, 90°). We calculated two
angles to describe the position of the upper arm related to
the thorax: (1) the angle of elevation (Figure 3(b)),
620
JRRD, Volume 47, Number 7, 2010
Figure 3.
Representations of calculated angles to express position of elbow and shoulder in accordance with recommendations of International Society of
Biomechanics: (a) elbow angle (), (b) angle of elevation (), and (c) plane of elevation ().
defined as the angle between the upper arm and trunk
(upper arm parallel with thorax, 0°; upper arm parallel
with horizontal; 90°), and (2) the plane of elevation (Fig-
ure 3(c)), defined as the angle between the thorax and the
upper arm in the transversal plane (arm extended forward,
0°; arm extended to the lateral, –90°).
For the maximal reaching task, we compared the
maximum elbow extension with and without gravity
compensation. To quantify the differences between the
reach and retrieval task with and without gravity com-
pensation, we derived joint rotations (in degrees) of the
angles just mentioned and parameters of the movement
cycles (mean duration of one movement cycle, number of
repetitions within 30 seconds). Cycle parameters were
averaged over all movement cycles within a series; the
first two cycles were excluded for analysis. A movement
cycle consisted of two parts, namely reach (maximum to
minimum elbow angle) and retrieval (minimum to maxi-
mum elbow angle).
Electromyography
Bipolar sEMG of eight superficial muscles (descend-
ing parts of the trapezius, anterior and posterior parts of
the deltoid, pectoralis major, long head of the biceps,
long head and lateral head of the triceps, and latissimus
dorsi) was recorded with circular, wet gel, silver/silver-
chloride electrodes (ARBO, type S93SG, Tyco/Health-
care Deutschland; Neustadt/Donau, Germany) at a sam-
ple frequency of 1,000 Hz. Electrode placement, skin
preparation, and recording protocol were in accordance
with the SENIAM guidelines [19].
sEMG signals were synchronized with the marker
trajectories (Figure 4). The time axis was normalized
from 0 to 100 percent: reach 0 to 50 percent and retrieval
51 to 100 percent.
We converted the band-pass filtered sEMG signals to
smooth rectified sEMG using a second-order Butterworth
filter with frequency at 25 Hz. To visualize the differ-
ences in smooth rectified sEMG, we plotted movement
trajectories (averaged data over all cycles) for two subjects
with and without gravity compensation plotted in the same
graph (Figure 5). Changes in the amplitude of muscle
activation during movements with gravity compensation
were expressed as a percentage of the change of the area
under the curve of the same movement without gravity
compensation. The area under the curve is calculated as
the integral of the smooth rectified sEMG.
Timing of muscle activation was analyzed visually.
The primary investigator assessed the sEMG recordings,
and a coauthor with extensive experience in sEMG analysis
checked it.
Statistical Analysis
This study had an explorative character; therefore,
the effect of gravity compensation was described separately
for each individual subject. Because of the small sample
size and a heterogeneous population, a Wilcoxon signed
rank test was performed and the median or ranges were
 621

KLOOSTERMAN et al. Gravity compensation in cervical SCI

Figure 4.
Elbow and shoulder joint angles (°) during 15 s repetitive reach and retrieval tasks with Freebal, performed by subject with identification number 2,
simultaneously displayed with smooth rectified surface electromyography values (microvolt) of eight measured muscles.
622

JRRD, Volume 47, Number 7, 2010

Figure 5.
Mean muscle activation pattern of reach and retrieval task with and without Freebal (Fb). Conditions with (dotted line) and without (solid line)
gravity compensation were plotted in same graph. Smooth rectified surface electromyography (sEMG) (microvolt) of eight measured muscles and
corresponding joint angles (°) was plotted against average movement cycle, divided into reach (1%–50%) and retrieval (51%–100%). (a) Activation
patterns of subject with identification (ID) number 9. Amplitude of sEMG of antigravity muscles decreased with use of Fb, except descending part of
trapezius. (b) Activation pattern of subject with ID number 2. Amplitude of sEMG of antigravity muscles decreased and amplitude of sEMG in
triceps increased with use of Fb.

found. From the Wilcoxon test, the test statistic T (smallest
of the two sums of ranks), its significance (p), and the
effect size (r) were reported.
RESULTS
Subjects
A complete data set was available for nine partici-
pants. The physical characteristics of each of the nine
subjects are displayed in Table 1.
Kinematics
Movement parameters are presented in Table 2. During
the maximal reaching task with and without gravity com-
pensation, the maximum elbow angle was significantly
lower with gravity compensation (median 33.3°) than
without gravity compensation (median 29.4°), T = 2, p =
0.021, r = –0.77.
During the reach and retrieval task with gravity com-
pensation, all subjects showed decreased elbow extension
(range 0.1°–11.0°). At the shoulder joint, seven subjects
 623

KLOOSTERMAN et al. Gravity compensation in cervical SCI

Table 1.
Physical characteristics of subjects (N = 9).
Subject ID Number
Variable
1 2 3 4 5 6 7 8 9
Sex M M M M F M F M M
Age (yr) 28 55 47 59 39 40 26 53 36
Time Since Injury (mo) 58 29 282 209 66 221 161 170 198
Motor Level of Lesion C5 C6 C6 C6 C5 C5 C6 C6 C7
ASIA Impairment Scale A C B D B C A A A
Measured Arm R L R L R L L L R
ASIA Motor Score of Measured Arm (MRC score)
C5 (elbow flexors) 4 5 5 5 4 5 5 5 5
C6 (wrist extensors) 0 4 4 4 5 2 5 5 5
C7 (elbow extensors) 0 2 0 5 4 4 0 0 3
C8 (finger extensors) 0 0 0 4 1 0 0 0 0
T1 (finger abductors) 0 0 0 4 0 0 0 0 0
ASIA = American Spinal Injury Association, C = cervical (fifth to eighth vertebra), F = female, ID = identification, L = left, M = male, MRC = Medical Research
Council, R = right, T = thoracic.

had decreased plane (0.3°–6.9°) and six subjects had
reduced angle of elevation (0.1°–15.1°). The movement
times increased in four subjects (range 0.1–0.4 s),
decreased in two subjects (0.2–0.4 s), and remained the
same in three subjects. None of these parameters differs
significantly between movements performed with and
without gravity compensation (elbow extension: T = 2, p =
0.214, r = –0.41; shoulder plane of elevation: T = 3, p =
0.767, r = –0.10; shoulder angle of elevation: T = 4, p =
0.515, r = –0.22; and cycle duration: T = 2, p = 0.484, r =
–0.23).
Electromyography
Based on the plotted smooth rectified sEMG (Figure 5)
and calculated differences (in terms of percentage) in the
areas under the curves (Table 2), we made three observa-
tions:
1. With gravity compensation, the amplitude of the
sEMG decreased especially in the antigravity muscles.
In six subjects, amplitude of the sEMG decreased in
the descending part of the trapezius (range 17.5%–
60.6%) and increased in three subjects (4.1%, 6.5%,
and 59.7%). In all subjects, amplitude of the sEMG
was decreased in the posterior part of deltoid (range:
12.8%–54.1%), the anterior part of deltoid (17.4%–
73.6%), and the long head of biceps (22.9%–80.0%).
2. In four subjects (identification numbers 1, 3, 7, and 8)
without triceps activity (Medical Research Council
[MRC] score of 0), sEMG activity was recorded dur-
ing flexion of the elbow.
3. In three of the five subjects with active triceps function
(MRC score of at least 2), the amplitude of sEMG in
the long head of triceps increased (25.2%, 1.2%, and
16.9%) and decreased in the other two subjects (16.4%
and 56.6%). On a group level, a significant difference
between the conditions with and without gravity com-
pensation was found for the following muscles:
descending part of trapezius during reach: T = 1, p =
0.038, r = –0.69; posterior part of deltoid during reach: T =
1, p = 0.015, r = –0.81, and during retrieval: T = 0, p =
0.008, r = –0.89; and anterior part of deltoid and long
head biceps for reach as well as retrieval: T = 0, p =
0.008, r = –0.89.
Within subjects, the timing of muscle activation did
not change visibly with gravity compensation. With
respect to the patterns of timing between subjects, we
found various different patterns. Some alternating activa-
tion patterns were found between agonists and antagonists.
All subjects with at least some triceps function showed
an alternating activation pattern between the long head of
biceps and triceps (Figure 6(a)). We found a simulta-
neous activation pattern in four subjects between the acti-
vation of the anterior and posterior parts of the deltoid
muscle (Figure 6(b)) and in six subjects between the ante-
rior part of deltoid and pectoral muscles (Figure 6(c)).
Furthermore, the descending part of the trapezius
was used in various different patterns. In one subject, an
alternating activation pattern between the anterior and
posterior parts of the deltoid occurred, and in another
subject, an alternating activation pattern between the
624

JRRD, Volume 47, Number 7, 2010

Table 2.
Influence of gravity compensation on kinematic parameters during maximal reaching task on kinematic and surface electromyography (sEMG)
parameters during reach and retrieval task of participants (N = 9) with and without Freebal (Fb).
Subject ID Number
Task Fb
1 2 3 4 5 6 7 8 9
Maximal Reach
No 47.9 19.0 47.5 36.9 22.3 24.3 21.1 46.2 29.4
Elbow Angle (°)
Yes 52.2 20.8 47.4 36.1 25.3 28.8 23.6 49.4 33.3
Reach and Retrieval
Angles of Elbow and Shoulder at Target Dot (°)
No 47.5 29.0 47.8 54.9 40.4 25.7 43.1 56.7 36.6
Elbow Angle
Yes 57.5 29.4 58.8 64.0 47.5 32.9 43.2 65.3 39.6
No –71.6 –42.7 –50.0 –66.4 –49.3 –39.8 –43.3 –56.7 –54.0
Plane of Elevation
Yes –64.7 –40.3 –49.7 –66.7 –48.4 –50.5 –41.4 –54.7 –51.2
No 30.6 23.0 37.8 51.3 29.1 27.2 44.6 44.7 40.4
Angle of Elevation
Yes 15.5 30.3 34.2 40.5 31.8 32.2 35.3 44.6 39.5
Parameters of Movement Cycles
No 2.7 2.1 1.4 2.4 1.3 1.6 1.5 2.9 1.3
Cycle Duration (s)
Yes 3.1 2.2 1.4 2.0 1.3 1.6 1.6 2.7 1.4
No 11 14 21 13 23 19 20 10 23
Repetitions (n in 30 s)
Yes 10 14 21 15 23 19 19 11 21
sEMG Parameters: Change of Area Under Curve (%)*
Descending Part of Trapezius — –17.5 –41.2 4.1 –60.6 –46.8 –48.3 59.7 6.5 –19.3
Posterior Part of Deltoid — –43.0 –54.1 –29.5 –16.0 –47.2 –26.1 –32.9 –26.3 –12.8
Anterior Part of Deltoid — –54.5 –31.9 –17.4 –73.6 –27.0 –33.2 –39.8 –30.6 –31.6
Pectoralis Major — –30.4 –51.2 –14.7 –40.6 –38.1 –11.3 –4.2 –39.0 22.2
Long Head Biceps — –57.0 –78.0 –42.9 –80.0 –41.6 –52.7 –44.4 –26.4 –22.9
Lateral Head Triceps — –55.7 46.6 –37.3 –24.8 –10.7 115.5 –43.4 –29.6 –13.8
Long Head Triceps — –47.2 25.2 –37.7 1.2 –16.4 –56.6 12.1 –24.5 16.9
Latissimus Dorsi — –17.5 –6.6 –13.1 –28.6 –23.5 –17.8 –5.5 45.4 –4.8
*Negative value means decrease in area under curve during movement with Fb, compared with same movement without Fb.
ID = identification.

posterior parts of the deltoid solely. These two combina-
tions were also observed in a simultaneous pattern: in one
subject, the descending part of the trapezius and anterior
and posterior parts of the deltoid were simultaneously
activated, and in another subject, the descending part of the
trapezius was activated with the posterior deltoid solely.
DISCUSSION
The objective of the present study was to study the
influence of gravity compensation on kinematics and
sEMG characteristics of the upper limb during goal-
directed movements of subjects with a CSCI.
With gravity compensation, most of the subjects
showed less elbow extension and movement execution
closer to the midline. Based on previous studies with
stroke patients, one can expect that gravity compensation
increases range of motion of the upper limb [12,14] because
of the positive effect on pathological muscle synergies
between shoulder abduction and elbow flexion [14]. In
patients with a CSCI, this pathological coupling does
not occur. However, an effect on kinematics is expected
because less muscle force is necessary to overcome

Figure 6.
Examples of muscle activation patterns and corresponding elbow
angle: (a) Alternating activation pattern between biceps and triceps
(subject identification [ID] number 2), (b) simultaneous pattern
between anterior and posterior part of deltoid (subject ID number 3),
and (c) simultaneous pattern between anterior deltoid and pectoralis
major (subject ID number 11).
gravity. A larger part of the muscle force could be used to
perform goal-directed movements, possibly leading to
increased elbow extension during maximal reaching and
more repetitions during reach and retrieval. However, the
625
KLOOSTERMAN et al. Gravity compensation in cervical SCI
results of this study showed less elbow extension during
maximal reaching with gravity compensation in seven
subjects. During reach and retrieval with gravity compen-
sation, all subjects showed less elbow extension and, in
eight subjects, a decrease in shoulder angle and/or plane
of elevation.
Plausible explanations could be given for these
results. First, subjects with a CSCI who have a lack of tri-
ceps function use their anterior part of the deltoid and
upper pectoral muscles to produce an isometric extension
torque in their elbow [20] or make a trick movement with
their shoulder muscles to achieve passive elbow exten-
sion [21]. They use gravity to maintain the arm in exten-
sion below the horizontal plane [22] and to perform a
passive elbow extension with a trick movement. In both
compensation strategies, gravity is used to maintain
elbow extension. Therefore, movement execution with
gravity compensation might decrease elbow extension.
Second, during goal-directed movements without gravity
compensation, subjects use a large part of the preserved
muscle force to hold the arm against gravity. If the pri-
mary agonists alone are not capable of generating the
required anteflexion and extension torques, additional
agonist muscles are recruited [23]. For example, the mid-
dle part of the deltoid might contribute to lift the arm, if
the anterior part of the deltoid cannot generate enough
force. The middle part of the deltoid also has an abduc-
tion function that can result in a reaching movement not
truly in the sagittal plane [23]. Third, because of a
decreased plane of elevation, the hand moves more in a
direct line to the target dot. If the arm is extended closer
to the midline, less elbow extension and angle of eleva-
tion are necessary to reach the target dot. Finally, with
gravity compensation, the pectoral muscles can move the
arm more easily to a position in front of the patient
because the weight of the arm is counteracted.
The results of the sEMG data during the reach and
retrieval task showed a decrease in sEMG activity during
movements with the use of the Freebal, particularly in
muscles that counteracted gravity, while timing remained
unaffected. The results confirmed our presumption based
on previous studies with nondisabled elderly [10] and
stroke patients [11,13,15] that also showed a decreased
sEMG in antigravity muscles and unaffected timing.
Remarkably, despite subjects with an MRC score of 0 in
the triceps, sEMG activity was seen mainly during elbow
flexion. A plausible explanation for this sEMG activity is
stretch or cocontraction. In the sEMG signal, however,
626
JRRD, Volume 47, Number 7, 2010
one cannot differentiate between activity because of
stretch and voluntary motor activity [24].
A large variety in muscle activation patterns was
seen between subjects because of heterogeneity of the
study population. After a CSCI, the functional anatomy
of the upper limb had to be redefined. Muscle synergies
as seen in nondisabled subjects are often inappropriate
for subjects with a CSCI [20]. The central nervous sys-
tem is challenged to use a motor strategy to adjust to the
new functional anatomy and biomechanics, with a
reduced repertoire of innervated muscles to deal with the
mechanics [21], leading to different movement patterns
between subjects with a CSCI [20].
To our knowledge, our study was the first explorative
study about the effect of gravity compensation on kine-
matics and sEMG in subjects with a CSCI. Another type
of arm support by subjects with a CSCI was studied by
Atkins et al. [7]. They reported about the effect of mobile
arm support on ADLs. Based on Delphi questionnaires,
they concluded that some ADLs were possible with the
use of a mobile arm support, which without the use of
such a device, patients with very weak biceps and deltoid
muscles were unable to perform.
Besides being used for compensating lost functions,
gravity compensation can be used for training purposes.
Further studies should be performed with a larger popula-
tion because of the small effect size, especially on kine-
matic parameters, and should be able to test the following
hypotheses: (1) patients with an MRC score of at least 2
in the triceps muscle can train their primary agonists of
the shoulder and elbow in goal-directed movements more
intensively and, (2) for patients without active triceps
function (MRC score of 0 or 1), gravity compensation
may not seem useful to train extension movements
because they perform these movements with the use of
gravity. However, gravity compensation might be benefi-
cial for training muscles required to cross the midline or
to perform bimanual tasks. Also, the influence of gravity
compensation on the patients’ ability to stabilize the
shoulder in a certain position would be an interesting
parameter.
CONCLUSIONS
This explorative study showed that gravity compen-
sation influenced the kinematics and amplitude of the
sEMG of the upper limb during goal-directed movements
in CSCI. A larger study is needed to firmly conclude
whether training with gravity compensation is clinically
relevent.
ACKNOWLEDGMENTS
Author Contributions:
Study concept and design: G. J. Snoek, M. Kouwenhoven, A. V. Nene,
M. J. A. Jannink.
Recruitment of study population: G. J. Snoek, M. Kouwenhoven, 
A. V. Nene.
Data collection and analysis: M. G. M. Kloosterman, 
M. Kouwenhoven.
Interpretation of data: M. G. M. Kloosterman, G. J. Snoek, 
M. Kouwenhoven, A. V. Nene, M. J. A. Jannink.
Drafting of manuscript: M. G. M. Kloosterman.
Critical revision of manuscript: G. J. Snoek, M. Kouwenhoven, 
A. V. Nene, M. J. A. Jannink.
Financial Disclosures: The authors have declared that no competing
interests exist.
Funding/Support: This material was based on work supported by the
Ministry of Economic Affairs (EZ) Overijssel and Gelderland, the
Netherlands, grant 1-5160. The study sponsor was not involved in any
aspect of this research.
Additional Contributions: Ms. Kloosterman is now only affiliated
with Roessingh Research and Development, Enschede, the 
Netherlands.
Institutional Review: The study was approved by the local medical
ethics committee. Subjects provided written informed consent before
being admitted to the study.
Participant Follow-Up: The authors do not plan to inform partici-
pants of the publication of this study.
REFERENCES
1. Kirshblum S, Ho CH, House JG, Druin E, Nead C, Drastal
S. Rehabilitation of spinal cord injury. In: Kirshblum S,
Campagnolo DI, DeLisa JA, editors. Spinal cord medicine.
Philadelphia (PA): Lippincott Williams & Wilkins; 2001.
p. 275–98.
2. Snoek GJ, IJzerman MJ, Hermens HJ, Maxwell D, Biering-
Sorensen F. Survey of the needs of patients with spinal cord
injury: Impact and priority for improvement in hand func-
tion in tetraplegics. Spinal Cord. 2004;42(9):526–32.
[PMID: 15224087] 
DOI:10.1038/sj.sc.3101638
3. Jacobs PL, Nash MS. Exercise recommendations for indi-
viduals with spinal cord injury. Sports Med. 2004;34(11):
727–51. [PMID: 15456347] 
DOI:10.2165/00007256-200434110-00003
4. Kloosterman MG , Snoek GJ, Jannink MJ. Systematic
review of the effects of exercise therapy on the upper

extremity of patients with spinal-cord injury. Spinal Cord.
2009;47(3):196–203. [PMID: 18825160] 
DOI:10.1038/sc.2008.113
5. Schmidt RA, Lee TD. Condition of practice. In: Schmidt RA,
Lee TD, editors. Motor control and learning: A behavioural
emphasis. 4th ed. Champaign (IL): Human Kinetics; 2005.
p. 321–63.
6. Powers SK, Howley ET. The physiology of training: Effect
on VO2 max, performance, homeostasis, and strength. In:
Powers SK, Howley ET, editors. Exercise physiology, the-
ory and application to fitness and performance. 5th ed.
New York (NY): McGraw-Hill; 2004. p. 249–75.
7. Atkins MS, Baumgarten JM, Yasuda YL, Adkins R, Waters
RL, Leung P, Requejo P. Mobile arm supports: Evidence-
based benefits and criteria for use. J Spinal Cord Med. 2008;
31(4):388–93. [PMID: 18959356]
8. Prange GB, Jannink MJ, Groothuis-Oudshoorn CG, Her-
mens HJ, IJzerman MJ. Systematic review of the effect of
robot-aided therapy on recovery of the hemiparetic arm
after stroke. J Rehabil Res Dev. 2006;43(2):171–84. 
[PMID: 16847784] 
DOI:10.1682/JRRD.2005.04.0076
9. Kwakkel G, Kollen BJ, Krebs HI. Effects of robot-assisted
therapy on upper limb recovery after stroke: A systematic
review. Neurorehabil Neural Repair. 2008;22(2):111–21.
[PMID: 17876068] 
DOI:10.1177/1545968307305457
10. Prange GB, Kallenberg LA, Jannink MJ, Stienen AH, Van
der Kooij H, IJzerman MJ, Hermens HJ. Influence of gravity
compensation on muscle activity during reach and retrieval
in healthy elderly. J Electromyogr Kinesiol. 2009;19(2):
e40–49. [PMID: 17911029] 
DOI:10.1016/j.jelekin.2007.08.001
11. Prange GB, Jannink MJ, Stienen AH, Van der Kooij H, IJzer-
man MJ, Hermens HJ. Influence of gravity compensation
on muscle activation patterns during different temporal
phases of arm movements of stroke patients. Neurorehabil
Neural Repair. 2009;23(5):478–85. [PMID: 19190089]
DOI:10.1177/1545968308328720
12. Stienen AH, Van der Helm FC, Prange GB, Jannink MJ,
Van der Kooij H. Effects of gravity compensation on the
range-of-motion of the upper extremities in robotic rehabili-
tation after stroke. In: Proceedings of the International
Shoulder Group (ISG). 2006 Oct 9–10; Chicago, IL; [about
4 screens]. Available from: http://www.arnostienen.net/
articles/stienen06.pdf/.
13. Prange GB, Stienen AH, Jannink MJ, Van der Kooij H, IJzer-
man MJ, Hermens HJ. Increased range of motion and
decreased muscle activity during maximal reach with grav-
ity compensation in stroke patients. In: Proceedings of the
2007 IEEE 10th International Conference on Rehabilitation
Robotics; 2007 Jun 13–15; Noordwijk aan Zee, the Nether-
627
KLOOSTERMAN et al. Gravity compensation in cervical SCI
lands. Piscataway (NJ): IEEE Press. p. 467–71.
DOI:10.1109/ICORR.2007.4428467
14. Dewald J, Yao J. The effect of generating anti-gravity
shoulder torques on upper limb discoordination following
hemiparetic stroke. In: Proceedings of the ISB XXth Con-
gress—ASB 29th Annual Meeting; 2005 Jul 31–Aug 5.
Cleveland, OH. International Society of Biomechanics. 973 p.
Available from: http://isbweb.org/o/content/view/280/1/.
15. Jannink MJ, Prange GB, Stienen AH, Van der Kooij H,
Kruidbosch JM, IJzerman MJ, Hermens HJ. Reduction of
muscle activity during repeated reach and retrieval with
gravity compensation in stroke patients. In: Proceedings of
the 2007 IEEE 10th International Conference on Rehabili-
tation Robotics; 2007 Jun 13–15; Noordwijk aan Zee, the
Netherlands. Piscataway (NJ): IEEE Press. p. 472–76.
DOI:10.1109/ICORR.2007.4428468
16. Maynard FM Jr, Bracken MB, Creasey G, Ditunno JF Jr,
Donovan WH, Ducker TB, Garber SL, Marino RJ, Stover
SL, Tator CH, Waters RL, Wilberger JE, Young W. Interna-
tional standards for neurological and functional classifica-
tion of spinal cord injury. American Spinal Injury
Association. Spinal Cord. 1997;35(5):266–74. 
[PMID: 9160449] 
DOI:10.1038/sj.sc.3100432
17. Stienen AH, Hekman EE, Van der Helm FC, Prange GB,
Jannink MJ, Aalsma AM, Van der Kooij H. Freebal: Dedi-
cated gravity compensation for the upper extremities. In:
Proceedings of the 2007 IEEE 10th International Confer-
ence on Rehabilitation Robotics; 2007 Jun 13–15; Noord-
wijk aan Zee, the Netherlands. Piscataway (NJ): IEEE
Press. p. 804–8. DOI:10.1109/ICORR.2007.4428517
18. Wu G, Van der Helm FC, Veeger HE, Makhsous M, Van
Roy P, Anglin C, Nagels J, Karduna AR, McQuade K,
Wang X, Werner FW, Buchholz B; International Society of
Biomechanics. ISB recommendation on definitions of joint
coordinate systems of various joints for the reporting of
human joint motion—Part II: Shoulder, elbow, wrist and
hand. J Biomech. 2005;38(5):981–92. [PMID: 15844264]
DOI:10.1016/j.jbiomech.2004.05.042
19. Hermens HJ, Freriks B, Merletti R, Stegeman D, Blok J,
Tau G, Disselhorst-Klug C, Hagg G. SENIAM: European
recommendations for surface electromyography. Results of
the SENIAM project. 2nd ed. Enschede (the Netherlands):
Roessingh Research and Development; 1999.
20. Marciello MA, Herbison GJ, Cohen ME, Schmidt R.
Elbow extension using anterior deltoids and upper pectorals
in spinal cord-injured subjects. Arch Phys Med Rehabil.
1995;76(5):426–32. [PMID: 7741612] 
DOI:10.1016/S0003-9993(95)80571-0
21. Koshland GF, Galloway JC, Farley B. Novel muscle pat-
terns for reaching after cervical spinal cord injury: A case
for motor redundancy. Exp Brain Res. 2005;164(2):133–47.
628
JRRD, Volume 47, Number 7, 2010
[PMID: 16028034] 
DOI:10.1007/s00221-005-2218-9
22. Acosta AM, Kirsch RF, Van der Helm FC. Three-dimen-
sional shoulder kinematics in individuals with C5–C6 spinal
cord injury. Proc Inst Mech Eng H. 2001;215(3):299–307.
[PMID: 11436273] 
DOI:10.1243/0954411011535894
23. McCrea PH, Eng JJ, Hodgson AJ. Saturated muscle activa-
tion contributes to compensatory reaching strategies after
stroke. J Neurophysiol. 2005;94(5):2999–3008. 
[PMID: 16014786] 
DOI:10.1152/jn.00732.2004
24. Moritani T, Stegeman D, Merletti R. Basic physiology and
biophysics of EMG signal regeneration. In: Merletti R, Parker
PA, editors. Electromyography. physiology, engineering,
and noninvasive applications. New York (NY): John Wiley
& Sons; 2004. p. 1–25. DOI:10.1002/0471678384
Submitted for publication February 9, 2010. Accepted in
revised form June 1, 2010.
This article and any supplementary material should be
cited as follows:
Kloosterman MGM, Snoek GJ, Kouwenhoven M, Nene
AV, Jannink MJA. Influence of gravity compensation on
kinematics and muscle activation patterns during reach
and retrieval in subjects with cervical spinal cord injury:
An explorative study. J Rehabil Res Dev. 2010;47(7):
617–28. 
DOI:10.1682/JRRD.2010.02.0014